Air based gas phase reactions which use metal oxide redox catalysts are used in chemical synthesis of acrylic acid, acrylonitrile, formaldehyde, maleic anhydride, acrolein, isophthalonitrile, nicotinonitrile and phthalic anhydride. A typical redox catalyst is vanadium-phosphorus, though others are well known in the art.
In the design of existing reactors both production and yield are taken into consideration. In this regard, it is recognized that there is a trade-off between production and yield such that parameters which provide a high level of production may, in fact, have the effect of decreasing product yield. For example, in order to increase production from an existing reactor, reactant feed rates must be increased; however, this has negative side-effects. Typically this procedure lowers the oxygen-to-feed ratio because air compressor and/or pressure drop limitations do not allow for an increase in the air flow rate.
Due to this lower oxygen-to-feed ratio, the partial pressure of oxygen in the reactor atmosphere may become insufficient to reoxidize the metal oxide catalyst which then becomes over-reduced and, eventually, deactivated. The net result is that product yields are depressed. Redox catalyst over-reduction also leads to a shortening of catalyst lifetime because the reduced form of these catalysts is relatively unstable.
The basic mechanism behind redox catalyst over-reduction can be understood by examining the following reactions which are applicable for any gas phase partial oxidations performed with metal oxide redox catalysts.
1) organic reactant+oxidized catalyst.fwdarw.product +reduced catalyst
2) reduced catalyst+oxygen.fwdarw.oxidized catalyst
As can be seen above, as the organic reactant reacts, the catalyst is reduced (Reaction 1). In order for the catalyst to be returned to its active oxidized state, it must be re-oxidized by gas phase oxygen (Reaction 2). If one has too much organic reactant and not enough oxygen, as when there is a high reactant feed rate, too much catalyst remains in the reduced state, and the catalyst is considered over-reduced.
As indicated above an over-reduced catalyst will be deactivated relative to the oxidized state. This deactivation is due to a combination of the following effects: chemical transformation of an active component into a less active component; reduction of active catalyst surface area through particle sintering, and the volatilization and loss of an active component. These effects are generally related to the unstable nature of a reduced catalyst and result in depressed reaction yield (e.g. the amount of desired product produced) and catalyst lifetime.
Thus, typically, manufacturers have accepted either the lowering of yield and catalyst lifetime associated with operating with a low oxygen-to-feed ratio, or the reduction in production associated with operating with a high oxygen-to-feed ratio.
One solution to this problem has been to add a continuous flow of oxygen to the air entering a reactor in order to maintain the oxygen-to-feed ratio during periods of increased production. This "oxygen enrichment" improves the rate of reoxidation of the catalyst, ameliorates over-reduction and thus allows one to maintain product yield while increasing the reactant feed flow to the reactor. This use of oxygen enrichment is usually only applicable to fluid bed reactors because these reactors are typically able to handle the increased heat load brought about by the increased amount of reaction. In this process, oxygen is typically injected into the air feed line of a reactor.
Using oxygen enrichment in the manner described above is applicable only in a retrofit application when market conditions make increased production from an existing plant desirable. Typically, such increases in production will only be desired for a fraction of the plant's operating life. Unfortunately this creates an fluctuating demand for oxygen which is difficult and costly to supply.
For a fixed bed reactor, continuous oxygen enrichment can be employed to increase the oxygen-to-feed ratio at a fixed production level or feed flow rate. This oxygen enrichment improves the rate of reoxidation of the catalyst, ameliorates over-reduction and thus allows one to increase product yield while maintaining the reactant feed flow to the reactor. Unfortunately, continuous oxygen enrichment is generally not economical as the savings associated with the yield increase are not enough to pay for the additional oxygen required.
For fixed bed reactors, the amount of oxygen added is usually between 1 and 3 vol. % of the total volume of all gases in the reaction atmosphere, as above this level there is no longer an improvement in yield. By the term "reaction atmosphere" we mean the total amount of all gases entering the reactor. If this oxygen were added to the air stream, this addition would result in a total oxygen concentration of about 22-24 vol. % in the air stream, or 1-3 vol. % enrichment. By the terms "volume % enrichment" or "% enrichment" we mean the difference between the oxygen vol. % in air and the oxygen vol. % in the mixture that would result if all the oxygen were added to the air stream.
It should noted that the oxygen concentration in the total volume of all the gases in the reactor is slightly less when compared to the oxygen concentration in the air stream-oxygen mixture, because the amount of oxygen is diluted by gaseous organic reactant which is present in an amount between 1 vol. % and 2 vol. % in fixed beds. The dilution factor is much greater with fluidized bed reactors as the amount of gaseous organic reactant is much higher. For example, the entering feed concentration, which includes ammonia in ammoxidation reactions, ranges from 4 vol. % for maleic anhydride to approximately 17 vol. % for acrylonitrile.
Several laboratory experiments have been conducted with metal oxide systems that vary the oxygen-to-feed ratio by cycling the reactant feed flow (Saleh-Ahlamad, 1992; Fiolitakis, 1983; Silveston, 1985). The reactant feed flow is varied either by pulsing the reactant feed on and off or at relatively high and low levels. Some selectivity improvement (e.g. how much of actual reacted starting material produces the desired product) has been noted in these experiments. However, in all but one example (Saleh-Ahlamad, 1992) the yield is lowered because of the reduction in conversion (e.g. the amount of starting material that actually reacts). Moreover, reactant feed cycling forces periodic operation of the entire plant, which adds to the complexity of the plant, and may actually reduce the overall performance of the plant, since most process equipment is designed to operate continuously.
Other laboratory experiments have alternately exposed metal oxide catalysts to reactant feed and to oxygen (Lang, 1989; 1991). This, in effect, is reactant feed and oxygen cycling. Some of these experiments have also included periodic flows of nitrogen to flush the catalyst. As with the reactant feed cycling experiments, while some selectivity increase was noted, product yield decreased. Further, such cycling increases the complexity required for plant operation.
Contractor, in U.S. Pat. No. 4,668,802, teaches a transport bed process for maleic anhydride which circulates the catalyst from a reaction zone where it is contacted with butane, to a stripping zone where the maleic anhydride is removed from the catalyst, and to a regeneration zone wherein the catalyst is contacted with an oxygen containing gas mixture. The oxygen and butane are never mixed together, thus effectively creating an alternating flow of oxygen and reactant feed with respect to the catalyst. This process enables high selectivities to be obtained while keeping throughput high.
However, transport bed technology is complex to design and operate and is not retrofitable. It is also difficult to produce the required attrition resistant catalyst. Finally, due to backmixing within the bed, the process is limited to chemicals capable of being produced in fluid beds. To date, the process has been applied only to maleic anhydride production.
As can be seen from the above discussion, under current processes one must accept either lower yields, lower production, or increased capital costs.